Through an API

Open and closed models answer the same suite-standard 1.0 through their APIs, ranked together. 1 ranked model.

#ModelPass rate
1DeepSeek V4.1 Flash 1017Open weights100%

Model leaderboard

84 models in chinese ranked by LMArena score · updated Nov 11, 2026

#ModelScore
1GLM-5.3Z.ai1535
2GLM-5.3-FlashZ.ai1528
3GLM-5.2Z.ai1522
4Kimi-K2.6Moonshot AI1521
5GLM-5.1Z.ai1516
6GLM-5Z.ai1512
7MiMo-V2.5-ProXiaomi MiMo1507
8DeepSeek-V4.1-FlashDeepSeek1505
9GLM-4.6Z.ai1500
10NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16NVIDIA1498
11Qwen3.5-397B-A17BQwen1498
12gemma-4-26B-A4B-itGoogle1495
13GLM-4.7Z.ai1494
14InklingThinking Machines1494
15Kimi-K2.5Moonshot AI1493
16Qwen3.8-27BQwen1493
17Hy3Tencent1490
18Qwen3-235B-A22B-Thinking-2507Qwen1482
19Qwen3.5-27BQwen1479
20gemma-4-31B-itGoogle1476
21MiMo-V2.5Xiaomi MiMo1470
22GLM-4.5Z.ai1465
23MiniMax-M3MiniMax1465
24Qwen3-235B-A22B-Instruct-2507Qwen1465
25MiMo-V2-FlashXiaomi MiMo1463
26Qwen3.5-122B-A10BQwen1463
27Inkling-SmallThinking Machines1462
28Qwen3-VL-235B-A22B-InstructQwen1462
29Qwen3-Next-80B-A3B-InstructQwen1460
30Mistral-Large-4-1T-A52BMistral AI1457
31Qwen3.5-35B-A3BQwen1457
32NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16NVIDIA1451
33LongCat-Flash-ChatMeituan LongCat1449
34Step-3.5-FlashStepFun1446
35Mistral-Large-3-675B-Instruct-2512Mistral AI1446
36Mistral-Medium-3.5-128BMistral AI1444
37Qwen3-VL-235B-A22B-ThinkingQwen1442
38MiniMax-M2.7MiniMax1439
39Hy3-previewTencent1435
40Qwen3-30B-A3B-Instruct-2507Qwen1433
41NVIDIA-Nemotron-3-Super-120B-A12B-BF16NVIDIA1432
42GLM-4.5-AirZ.ai1426
43Qwen3-235B-A22BQwen1424
44GLM-4.6VZ.ai1424
45granite-4.2-30bIBM Granite1423
46Ling-flash-2.0inclusionAI1419
47Muse-Glimmer-30Bmeta-models1410
48Qwen3-Next-80B-A3B-ThinkingQwen1406
49GLM-4.7-FlashZ.ai1403
50MiniMax-M2.5MiniMax1393
51gpt-oss-120bOpenAI1385
52NVIDIA-Nemotron-3-Nano-30B-A3B-BF16NVIDIA1385
53QwQ-32BQwen1378
54Trinity-Large-ThinkingArcee AI1373
55MiniMax-M2MiniMax1366
56NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16NVIDIA1365
57granite-4.2-8bIBM Granite1362
58Qwen3-32BQwen1357
59Qwen3-Coder-480B-A35B-InstructQwen1356
60INTELLECT-3Prime Intellect1355
61Trinity-Large-PreviewArcee AI1353
62Qwen3-30B-A3BQwen1347
63GLM-4.5VZ.ai1337
64granite-4.1-8bIBM Granite1336
65gpt-oss-20bOpenAI1315
66Olmo-3.1-32B-InstructAi21304
67granite-4.2-3bIBM Granite1279
68Llama-4-Maverick-17B-128E-InstructMeta Llama1278
69Qwen2.5-72B-InstructQwen1272
70Llama-4-Scout-17B-16E-InstructMeta Llama1255
71granite-4.0-h-smallIBM Granite1251
72Llama-3.1-405B-InstructMeta Llama1242
73Qwen2.5-Coder-32B-InstructQwen1222
74QwQ-32B-PreviewQwen1218
75Llama-3.3-70B-InstructMeta Llama1217
76Llama-3.1-70B-InstructMeta Llama1215
77Llama-3.1-8B-InstructMeta Llama1151
78granite-3.1-8b-instructIBM Granite1145
79granite-3.1-2b-instructIBM Granite1139
80Mixtral-8x22B-Instruct-v0.1Mistral AI1116
81granite-3.0-2b-instructIBM Granite1070
82granite-3.0-8b-instructIBM Granite1063
83Llama-3.2-3B-InstructMeta Llama1017
84Llama-3.2-1B-InstructMeta Llama959
Show the top 20

Score by size

Plot against
Best for its size

Scores: LMArena, CC BY 4.0. Measured on served versions, not local builds. A model with several served variants shows its best.